The New ChatGPT Images Is Here
OpenAI is rolling out a new ChatGPT Images experience powered by its latest image generation model, available to all ChatGPT users and in the API as GPT‑Image‑1.5. The release focuses on precision editing that preserves what matters, more expressive transformations, stronger instruction following, and improved text rendering—now with image generation up to 4× faster.
Table of Contents
Precision editing and creative transformations
The new model is designed to change exactly what you ask for—while keeping lighting, composition, and likeness consistent across inputs, outputs, and subsequent edits. That fidelity turns ChatGPT into a pocket creative studio for both practical edits and expressive reimaginings.
Editing: change only what you intend
- Add, remove, combine, blend, or transpose elements without losing the image’s essence.
- Preserve subjects’ identity and scene continuity across iterative edits (e.g., wardrobe try‑ons, hairstyle tests).
- Maintain global properties like lighting and composition while applying local changes.
Creative transformations: reframe concepts, keep key details
- Turn photos into movie posters, ads, ornaments, or paintings while retaining core identity cues.
- Manipulate layout and inject text elements while preserving readability and design balance.
- Explore stylistic presets directly from the new Images feature for prompt‑free experimentation.
Instruction following: more consistent compositional control
- Better adherence to complex, multi‑element prompts (e.g., structured grids, multi‑row object layouts).
- Improved relational accuracy—objects appear where and how you specify.
Text rendering: denser, smaller, more legible
- Stronger rendering of compact text blocks and small fonts.
- Suitable for posters, UI mockups, infographics, and document‑like assets.
New Images experience and API
OpenAI is introducing a dedicated Images space in the ChatGPT sidebar (mobile and web) to make exploration faster and simpler. It ships with preset filters, trending prompts, and an optional one‑time likeness upload so you can reuse your appearance across future creations without digging through your camera roll.
A creation space built for iteration
- Generate while previous images are still in progress to reduce wait time.
- Up to 4× faster rendering for quicker idea loops.
- Preset styles to spark inspiration without writing prompts.
GPT‑Image‑1.5 in the API
- Available now via the OpenAI API with the same preservation and editing upgrades.
- 20% cheaper image inputs and outputs versus GPT Image 1, enabling more iterations per budget.
- Better brand/logo consistency across edits—well‑suited for marketing, ecommerce catalogs, and design workflows.
- Try it in Playground, browse the gallery, or consult the prompt guide to refine your requests.
Availability and rollout
- The new Images model is rolling out to all ChatGPT users today; most users get the new Images experience now, with Business and Enterprise access coming later.
- The earlier ChatGPT Images version remains available as a custom GPT for continuity.
FAQ
What exactly is new in ChatGPT Images?
A new image model (GPT‑Image‑1.5) with precise editing, stronger instruction following, better text rendering, more expressive transformations, and faster generation.
How well does it preserve likeness and composition?
Significantly better than the initial version. It keeps lighting, framing, and identity consistent across edits, which is crucial for portraits, try‑ons, and brand visuals.
Can it handle dense, small text?
Yes—text rendering is improved, enabling more legible newspaper‑like layouts, infographics, and UI mockups.
What are the current limitations?
While improved across many cases, limitations persist in certain styles, multilingual outputs, and multi‑face scenes. Future iterations aim to narrow these gaps further.
Is it available for enterprises?
The new Images experience is rolling out to most users now; Business and Enterprise access is coming later.
How fast is generation?
Images now render up to 4× faster, and you can keep generating while previous images are in progress.
What about API pricing and use cases?
GPT‑Image‑1.5 is about 20% cheaper for image inputs/outputs than GPT Image 1. It’s well‑suited for ecommerce product variants, marketing graphics, and logo‑consistent campaigns.
Conclusion
ChatGPT Images with GPT‑Image‑1.5 marks a meaningful step forward: edit precisely without breaking composition, transform creatively without losing identity, render legible dense text, and iterate faster in a dedicated creation space. With lower API costs and better brand preservation, it also expands practical use across marketing, ecommerce, and creative tooling.
If you need AI photo editing for object removal or background replacement outside ChatGPT, you can try Imgezy (https://www.imgezy.com/).